sourdough fermentation or addition of organic acids or

Human and Clinical Nutrition

Sourdough Fermentation or Addition of OrganicAcids or Corresponding Salts to Bread ImprovesNutritional Properties of Starch in Healthy Humans1'2

HELENA G. M. LILJEBERG,3 CLAS H. LÃ–NNER* Ano IMGER M. E. BJÃ–RCK

Department of Applied Nutrition and Food Chemistry, Chemical Center,University of Lund, P.O. Box 124, S-221 00 Lund, Sweden and *Clas LÃ¶nnerAB,Ideon, S-223 70 Lund, Sweden

ABSTRACT Postprandial blood glucose and insulinresponses to barley bread containing organic acids orcorresponding salts were evaluated in healthy humansubjects. The satiety score and the rate and extent of invitro starch digestion were also studied. Lactic acid wasgenerated by use of a homofermentative starter cultureor added to the dough. In addition, products were bakedwith Ca-lactate, or with Na-propionate at two differentconcentrations. Consumption of the product baked witha high concentration of Na-propionate significantly lowered the postprandial blood glucose and insulinresponses, and significantly prolonged the duration ofsatiety compared with all other breads. When subjectsconsumed the breads baked with sourdough, lactic acidand Na-propionate, their glucose and insulin responseswere reduced compared with the wholemeal bread alone.The rate of in vitro amylolysis was reduced only byingestion of the breads containing lactic acid, suggestingthat the beneficial impact of Na-propionate on metabolicresponses and satiety was related to effects other than areduced rate of starch hydrolysis. All bread products hada similar concentration of in vitro resistant starch of1.3-2.1 g/100 g (starch basis). It is concluded thatsourdough baking and other fermentation processesmay improve the nutritional features of starch. Theresults also demonstrate that certain salts of organicacids may have metabolic effects. J. Nutr. 125:1503-1511, 1995.

INDEXING KEY WORDS:

â€¢sourdough fermentation â€¢starch hydrolysisâ€¢glycÃ©mieresponse â€¢satiety â€¢humans

In recent years it has become increasingly clearthat starchy foods have very different effects on postprandial blood glucose and insulin responses. Starchin legumes (Tovar et al. 1992), pasta (Granfeldt andBjÃ¶rck 1991) and products based on whole cerealgrains (Granfeldt et al. 1994) elicits low glucose andinsulin responses, whereas starch in potatoes (Collieret al. 1986) and most conventional bread products

(Jenkins et al. 1988) is rapidly digested and absorbed,thus giving rise to high metabolic responses. Foodswith slowly digested and absorbed carbohydrates areof special importance in the dietary management ofdiabetic patients (Brand et al. 1991). However, dietscharacterized by such foods have also been found toalter lipid and glucose metabolism in healthy subjects(Jenkins et al. 1987), as judged from reductions intotal cholesterol levels and improved glucosetolerance. In fact, dietary carbohydrates that cause arapid rise in postprandial insulin levels are extensively discussed as a risk factor for the developmentof metabolic diseases (Ducimetiere et al. 1980). Otheradvantageous effects of lente food items, registered innondiabetic subjects, include a prolonged duration ofsatiety (Haber et al. 1977, Holt et al. 1992) andphysical endurance during athletic performance(Thomas et al. 1991).

Because of the nutritive composition and low fatcontent of bread, an increased bread intake is advocated. Unfortunately, the starch in most breadproducts causes unfavorably high metabolicresponses. Attempts to reduce the glycÃ©mieresponseto bread have therefore been made. Previously, weshowed that cereal kernels included in bread (wheat,rye or barley) reduced the glycÃ©mieresponse to breadin healthy humans (Liljeberg and BjÃ¶rck 1994, Lil-jeberg et al. 1992). In fact, an inverse relationshipbetween the ratio of barley kernels and glycemia wasseen (Liljeberg and BjÃ¶rck 1994). The lowered metabolic response to these kernel-based products was

'Supported by grants from the Cerealia Foundation for Research

and Development (project no. 193).2The costs of publication of this article were defrayed in part by

the payment of page charges. This article must therefore be herebymarked "advertisement" in accordance with 18 USC section 1734

solely to indicate this fact.3To whom correspondence and reprint requests should be ad

dressed.

0022-3166/95 $3.00 Â©1995 American Institute of Nutrition.Manuscript received 29 August 1994. Initial review completed 3 October 1994. Revision accepted 1 December 1994.

1503

by guest on August 2, 2014

jn.nutrition.orgD

ownloaded from

http://jn.nutrition.org/

1504 LILIEBERG ET AL.

closely related to a lowered enzymic susceptibility ofstarch, as judged from studies in vitro (Liljeberg andBjÃ¶rck 1994, Liljeberg et al. 1992). To provide anenzymic barrier, the cereal kernels must retain theirintegrity in the finished bread, and an oat kernelbread with disrupted kernels did not improve theglucose response compared with a white referencebread (Liljeberg et al. 1992).

In contrast, despite disrupted rye kernels, a commercial sourdough fermented pumpernickel breadwas recently found to flatten the postprandial glucoseand insulin response curves in healthy subjects (Liljeberg and BjÃ¶rck 1994). Improved glucose tolerancemay result from the presence of organic acids thatmay delay gastric emptying (Hunt and Knox 1972) orinhibit amylase activity (Todesco et al. 1991) and,hence, reduce the rate of small intestinal starchuptake.

The aim of the present study was to evaluate thepossible influence of acids formed during sourdoughfermentation in bread on the postprandial glucose andinsulin responses in healthy subjects. Lactic acid wasgenerated by use of a homofermentative starterculture or was added to the dough. To investigate theimportance of pH per se, the corresponding salt, Ca-lactate, was included in one product. In addition, twobread products were baked with different concentrations of Na-propionate. In the baking industry, propi-onate is commonly used to inhibit mold and bacterialgrowth. Also studied were the rate and extent of invitro starch digestion. Finally, the bread productswere evaluated with respect to satiety and acceptability scores.

MATERIALS AND METHODS

Bread products. White wheat flour and whole-mealbarley flour were provided by Nord Mills AB (MalmÃ¶,

Sweden). All bread products were made from thesame basic recipe with 80% whole-meal barley flourand 20% white wheat flour. Two of the productscontained lactic acid. One was baked with sourdough,generated by use of a homofermentative starterculture, which formed mainly lactic acid. The otherwas baked with addition of lactic acid. Two productswere also baked with either the calcium salt of lacticacid, Ca-lactate, or the sodium salt of propionic acid,Na-propionate. The four bread products had the samecontent of organic acid or salt on a molar basis. Inaddition, a product with a high content of Na-propionate was included in the study. A whole-meal breadwith no additives (WMB) was used as a referenceproduct.

Recipes. To make sourdough, 3465 g water, 1540 gwhole-meal barley flour and 0.75 g starter culture(5-107 colony forming units/g flour) were mixed andstored for 20 h at 37Â°C.The starter culture used was a

homofermentative lactic acid bacteria, Lactobacillusplantarum strain Al (CÃasLÃ¶nner AB, Lund, Sweden)(Spicher and LÃ¶nner 1985). The main organic acidformed during fermentation was lactic acid (Table 1).

The basic recipe for the whole-meal bread consisted of 3280 g water, 2960 g whole-meal barleyflour, 740 g white wheat flour, 200 g yeast, 50 g NaCl,50 g sucrose and 37 g monoglycerides. The dough wasproofed for 50 min, divided into pieces of 600 g,followed by a second proofing for 20 min (38Â°C,75%humidity). The bread was baked at 200Â°Cfor 30 min.

Sourdough bread was made with 4810 g sourdoughstarter, 1480 g whole-meal barley flour, 740 g whitewheat flour and 200 g yeast. The remaining ingredients and baking procedures were the same as for thebasic recipe.

To make the other four bread products, the basicrecipe was used, to which was added 64 g of lacticacid (90 wt%), 95 g of Ca-lactate, 62 g of Na-propionate or 185 g of Na-propionate. The water content

TABLE 1

Composition of the bread products

Bread product

Acid StarchpH SÂ°' Lactic2 Propionic2 Ethanol Available Resistant Dietary fiber

mmol/100 g drywtWhole-mealbread(WMB)3Plus

sourdoughPluslacticacidPlus

Ca-lactatePlusNa-propionatePlus

Na-propionate, high concentration5.94.24.05.35.96.03.521.423.18.68.78.50.018.417.918.90.00.00.00.00.00.021.360.81.01.21.11.10.60.3g/100

g66.765.664.064.765.062.6drywt1.40.91.11.11.21.113.312.513.513.112.712.0

'SÂ°â€¢=acid equivalents, expressed as the amount of 0.1 mol/L NaOH consumed in mL/10 g bread |dry wt).2Minor amounts, 0.04-0.08 g/100 g acetic acid (dry wt basis) formed during baking is included.3WMB = whole-meal bread (whole-meal barley flour, white wheat flour, 80:20), reference product.


jn.nutrition.orgD

ownloaded from


SOURDOUGH OR ORGANIC SALTS AND NUTRITION OF STARCH 1505

was adjusted accordingly when lactic acid was added.Baking procedures were the same as for the basicwhole-meal bread.

Before being cut into slices, all bread products werestored at room temperature overnight. The crust wasremoved and three to four slices were wrapped inaluminum foil, put into plastic bags and stored at-20Â°C until used.

Chemical analysis. A portion from each bread (<0.8mm) was air dried and milled (Cyclotec, Tecator,Sweden) before analysis. The bread products wereanalyzed for potentially available starch (Holm et al.1986) and total dietary fiber (Asp et al. 1983). Further,total starch remnants in the fiber residues were determined enzymatically following solubilization inalkali (total in vitro resistant starch). Residual starchremnants were analyzed directly, omittingpretreatment in alkali. However, the thermostable a-amylase Termamyl was not included (SiljestrÃ¶m et al.1989). The fraction that required solubilization inalkali, presumably retrograded amylose, was then calculated as the difference between total starch andresidual starch remnants in the fiber residues.

The bread products were also characterized withrespect to pH and acid equivalents (SÂ°)(LÃ¶nner and

Preve-Ã€kesson 1988). Samples for measuring theamounts of lactic acid, propionic acid and ethanolwere prepared and stored as described by LÃ¶nner andPrevÃ©-Akesson (1988). The amount of lactic acid wasdetermined enzymatically (Boehringer Mannheim,GmbH Biochemica, Mannheim, Germany) and theamounts of acetic acid, propionic acid and ethanolwere measured on a gas Chromatograph (Varian 3400,VarÃan, Palo Alto, CA).

The composition of the bread products is shown inTable 1.

Blood glucose and insulin responses in healthysubjects. Eleven healthy volunteers (six women andfive men, age 26-48 y, with normal body mass indicesand without drug therapy) participated in the study.The bread products (159-161 g) were provided cor

responding to 50 g of available carbohydrates andserved with 8 g of butter and 20 g of cheese (17% fat,wet weight). In addition, 200 mL of water and 150 mLof coffee or tea were included in each meal. Theenergy content of the test meals was 1590 KJ. Valuesfor protein and fat in the bread products were basedon analysis of products with similar composition,described by Liljeberg and BjÃ¶rck (1994). Corresponding values for cheese and butter were obtainedfrom food tables. The subjects were served the testproducts as a breakfast in random order (six separateoccasions) after an overnight fast. They were asked toeat the meal over a 12-15 min period.

Finger prick capillary blood samples were takenbefore the meal (0) and at 30, 45, 70, 95, 120 and 180min after the meal for analysis of glucose, and after30, 45, 95 and 120 min for analysis of insulin. Blood

glucose concentration was determined with a glucoseoxidase peroxidase reagent and plasma insulin levelwas measured with an enzyme immunoassay kit(Boehringer Mannheim).

Approval of the study was given by the EthicsCommittee at the University of Lund, Lund, Sweden.

In vitro starch hydrolysis. The bread productswere tested in vitro to determine the rate of release ofstarch hydrolysis products following incubation withsalivary a-amylase, pepsin and pancreatic a-amylase(Granfeldt et al. 1992). Subjects rinsed their mouthswith tap water and subsequently chewed the breadproducts for 15 s (-15 times). They were told not toeat within l h before the experiment. All bread portions contained l g of starch and were given in arandomized order to six subjects. The products werethen expectorated into a beaker containing 50 mg ofpepsin (2000 FIP-U/g, Merck, Darmstadt, Germany) in6 mL of 0.05 mol/L Na,K-phosphate buffer (containing 0.4 g/L NaCl) adjusted to pH 1.5 with 2 mol/LHC1. Finally, the subjects rinsed their mouths with 5mL of Na,K-phosphate buffer (pH 6.9) for 60 s andexpectorated the rinsing solution into the beaker.

The contents were stirred and pH adjusted to 1.5.Each sample was incubated at 37Â°Cfor 30 min with

gentle mixing three times during incubation. The pHwas readjusted to 6.9 with NaOH before incubationwith porcine pancreatin a-amylase (A6255, SigmaChemical, St. Louis, MO). The enzyme (110 Sigmaunits) was dissolved in 10 mL of buffer, and 1 mL ofthis solution was added to the beaker. The samplewas adjusted to 30 mL with phosphate buffer, andtransferred to the dialysis tubing (13 cm strips,Spectra Por no. 2, width 45 mm, molecular weightcutoff 12,000-14,000). Each bag was incubated at37Â°Cfor 3 h in a beaker with phosphate buffer (800

mL). The beaker was placed in a stirred water bath.Every 30 min, aliquots (2 mL) from the dialysate wereremoved for analysis of reducing sugar content by the3,5-dinitro salicylic acid method (Hostettler et al.1951). A standard curve was prepared using maltose.In addition, a hydrolysis rate index was calculated as100 times the area under the curve (0-180 min) forthe product divided by the corresponding area obtained with white wheat bread chewed by the sameperson (Liljeberg and BjÃ¶rck 1994).

Satiety and acceptability scores. The satiety afterthe test meals was estimated numerically accordingto Haber et al. (1977). Assessments were done beforethe meal (0) and at 15, 45, 95, 120 and 180 min afterthe meal using a scoring system graded from -10, torepresent extreme hunger, to +10, to represent extreme satiety. The area under the satiety curve wascalculated above zero line.

During the satiety study, the subjects were alsorequested to assess the acceptability of the breads ona bipolar hedonic scale (Chapman and Wigfield 1970),in which -4 represents dislike extremely, Â±0


jn.nutrition.orgD

ownloaded from


1506 LILJEBERG ET AL.

represents neither like nor dislike and +4 representslike extremely.

Statistical methods. The results are expressed asmeans Â±SEM,and the statistical significance of differences were assessed by the Wilcoxon matched-pairsigned-ranks tests (Conover 1980). The SPSS/PC+ advanced statistics program (version 2.0, SPSS, Chicago,IL) was used. A value of P < 0.05 was consideredsignificant.

RESULTS

Postprandial blood glucose and insulin responses.Consumption of the breads with added Na-propionateresulted in significantly lower blood glucose concentrations (P < 0.05) at the initial postprandial phase (30min), than did consumption of the WMB (Fig. 1 andTable 2). The glucose concentration measured 30 minafter ingestion of the WMB plus high Na-propionatewas significantly lower than with the remainingproducts tested. When the areas under the curves inthe initial phase were calculated (0-45 min), significantly lower figures were noted as a result of consumption of the sourdough bread, and the breads withlactic acid or Na-propionate, compared with WMB(Table 3). However, the 0-95 and 0-120 min areasunder the curves did not differ significantly. Nolowering of postprandial glycemia was found as aresult of ingestion of the WMB plus Ca-lactate.

More pronounced differences were obtained inpostprandial insulin responses. Compared with WMB,lower insulin concentrations (P < 0.05) were found at30 min when subjects consumed the sourdough bread,and the WMB plus lactic acid or Na-propionate (Fig. 2and Table 4). Also, a significantly lower insulinincrement was observed following ingestion of WMBplus high concentration of Na-propionate, comparedwith all the remaining products. When subjects consumed the sourdough bread, and the breads withadded lactic acid or Na-propionate, the 0-45 minareas under the curves were significantly distin-

/NJS

0

0030TJ00iO

30 AS TO 95 120

Time <min>

ieo

FIGURE 1 Mean incremental blood glucose responses inhealthy subjects following ingestion of breakfast meals withdifferent bread products. Values are means, n - 11. Errorterms are given in Table 2.

guished from the area obtained with WMB (Table 5).The same differences persisted when calculating the0-95 min areas under the curves.

Rate of in vitro starch hydrolysis. The percentageof starch hydrolyzed within 180 min (mean Â±SEM)was 53.2 Â±1.0 for WMB and 48.6 Â±0.9, 46.8 Â±0.6,58.7 Â±1.6, 55.9 Â±0.3, 55.2 Â±1.0 for the sourdoughbread, WMB plus lactic acid, WMB plus Ca-lactate,WMB plus Na-propionate and WMB plus the highconcentration of Na-propionate, respectively. Whenthe hydrolysis rate index was calculated, significantlylower indices (P < 0.05) were found with the sourdough bread and the WMB plus lactic acid (86 and 81,respectively), compared with white wheat bread andthe WMB (Table 6). The hydrolysis rate index obtained with the breads with added Na-propionatewere not significantly distinguished from that withwhite wheat bread. In contrast, the indices found

TABLE 2

Incremental blood glucose responses in healthy humans after different bread meals1

Glucose response

30 min 45 min 70 min 95 min 120 min 180 min

Ammol/LWhole-mealbreadPlussourdoughPlus

lacticacidPlusCa-lactatePlusNa-propionatePlus

Na-propionate, high concentration2.1

Â±0.1a1.6Â±O.lab1.7Â±0.2ab1.8Â±0.1a1.4Â±0.2b1.0Â±0.2C1.5

Â±0.31.1Â±0.31.1Â±0.21.4Â±0.21.4Â±0.21.1Â±0.20.7

Â±0.2ab0.5Â±0.3a0.9Â±0.3ab0.4Â±0.2ab0.8Â±0.4b1.0Â±0.3ab0.3

Â±O.lab0.1Â±0.1a0.4Â±0.2ab0.0Â±0.1ab0.6Â±0.2b0.5Â±0.2ab0.1

Â±0.10.1Â±0.10.2Â±0.10.0Â±0.20.0Â±0.10.0Â±0.1-0.-0.-0.-o.:-0.-0.Â±

0.1Â±0.1Â±

0.1.Â±0.1Â±0.1Â±

0.1

'Values are means Â±SEM,n - 11. Values not sharing the same letters are significantly different (P < 0.05).


jn.nutrition.orgD

ownloaded from



TABLE 3

Blood glucose and postprandial blood glucose responses in healthy humans after different bread meals1

Whole-mealbreadPlussourdoughPlus


Na-propionate, high concentrationFasting

concentrationmmol/L4.4

Â±0.1a4.6Â±O.lab4.5Â±0.1a4.6Â±O.lb4.5Â±0.1a4.4Â±O.lab0-45

min57.4

Â±4.3a44.9Â±4.3bc47.8Â±4.6b51.1Â±4.1ab41.1Â±4.5bc30.5Â±3.8CArea

undercurve0-95

minmmol/L

min96.9

Â±12.874.1Â±15.088.7Â±12.681.1Â±9.687.0Â±15.575.9Â±11.60-120

min103.0

Â±13.678.6Â±15.197.2

Â±14.186.1Â±9.795.5Â±16.284.6Â±12.9

'Values are means Â±SEM,n = 11. Values not sharing the same letters are significantly different (P < 0.05).

with these products were significantly higher compared with WMB. Similarly, the WMB plus Ca-lactateshowed a significantly higher hydrolysis rate indexthan white wheat bread and WMB. No significantdifferences in hydrolysis rate index were found between white wheat bread and WMB.

In vitro resistant starch. The contents of in vitroresistant starch were similar in all bread products(0.9-1.4 g/100 g dry weight basis), corresponding to1.3-2.1 g/100 g starch basis (Table 1). The main indigestible fraction, or -50-65%, consisted of a starchfraction, presumably retrograded amylose, which required solubilization in alkali to render it availablefor the analytical amylases.

Satiety and acceptability scores. In the earlyphase, 45 min after the ingested meal, the two breadproducts with added Na-propionate were given highersatiety scores (P < 0.05) than the WMB product (Fig. 3and Table 7). For the bread plus high Na-propionate,this was also true at 95 min. In addition, at 120 min,significantly higher satiety scores were registeredwith WMB plus Ca-lactate and WMB plus Na-propionate, compared with WMB. When calculating thesatiety area under the curves (0-180 min), a signifi

cantly higher value was found with the bread bakedwith the high concentration of Na-propionate thanwith WMB (Table 8). The highest acceptability scoreswere given to the WMB, WMB plus Ca-lactate andWMB plus Na-propionate (Table 8). No significantdifferences (P < 0.05) were found between theseproducts. The sourdough bread, WMB plus lactic acidand WMB plus high Na-propionate were liked bysome, but also disliked by others, resulting in scoressignificantly lower than with WMB.

DISCUSSION

Results from this study indicate that consumptionof bread products containing lactic acid, whethergenerated during fermentation or added, reduced postprandial glucose and insulin responses in healthy subjects. Also, breads baked with Na-propionate improved the same metabolic indices. The largestreduction was obtained with the bread baked with ahigh concentration of Na-propionate. The area underthe blood glucose curve (45 min) was reduced by

TABLE 4

Incremental serum insulin responses in healthy humans after different bread meals1

Insulin response

30 min 45 min 95 min 120 min

Anmol/LWhole-mealbreadPlussourdoughPlus


Na-propionate, high concentration0.38

Â±0.05a0.29Â±0.06b0.25Â±0.03b0.32Â±0.05ab0.20Â±0.04ab0.12Â±0.03C0.32

Â±0.070.24Â±0.050.27Â±0.070.30Â±0.070.22Â±0.050.21Â±0.050.11

Â±0.04ab0.09Â±0.03a0.11Â±0.04ab0.07Â±0.02ab0.14Â±0.04b0.11Â±0.03ab0.05

Â±0.030.06Â±0.020.05Â±0.020.05Â±0.010.07Â±0.020.05Â±0.02



jn.nutrition.orgD

ownloaded from



TABLE5

Fasting serum insulin and postprandial serum insulin responses in healthy humans after different bread meals1

FastingconcentrationArea

undercurve0-45minnmol/LWhole-meal

breadPlusPlusPlusPlusPlussourdoughlactic

acidCa-lactateNa-propionateNa-propionate,

high concentration0.06

Â±0.05Â±0.05Â±0.06Â±0.05Â±0.05

Â±0.010.010.010.010.010.0110.98.47.69.46.04.3Â±Â±Â±Â±Â±Â±1.4a1.6bc12bc1.5abl.lbc1.0C0-95

minnmol/L21.7

Â±16.7Â±17.2Â±18.7Â±15.0Â±12.2

Â±min3.7a3.2bc41bc3.6ab2.6bc2.3C0-12023.718.519.320.117.714.2Â±Â±Â±Â±Â±Â±min4.53.74.83.93.02.4

'Values are means Â±SEM,n = 11. Values not sharing the same letters are significantly different (P < 0.05|.

-47%, compared with the reference product. This is

in agreement with Tedesco et al. (1991), who reporteda 48% reduction of the glucose area (120 min) with asimilar bread product. In the present study, the postprandial insulin areas were also evaluated. The insulin responses with WMB plus high Na-propionatewere very low, and the 45- and 95-min areas were

lowered by 61 and 44%, respectively.The sourdough bread, and the breads with added

lactic acid, Ca-lactate and Na-propionate were bakedwith the same content of organic acid or salt on amolar basis. Ingestion of the propionate bread (pH 5.9)reduced metabolic response similarly to the sourdough (pH 4.2) or lactic acid (pH 4.0) breads. Theglucose areas (45 min) were reduced by 17-28%, andthe effect on insulin responses was even more pronounced. Consumption of the sourdough bread, andthe breads with added Na-propionate and lactic acid

3

C

E

eM

o..Â«*

0.3

O.2

O.I

O 30 12O

"Time <min)

FIGURE 2 Mean incremental serum insulin responses inhealthy subjects following ingestion of breakfast meals withdifferent bread products. Values are means, n = 11. Errorterms are given in Table 4.

decreased the mean 45-min areas by 45, 23 and 30%,respectively, and the 95-min areas by 31, 23 and 21%,respectively. Although not statistically significant (P< 0.05), a similar tendency was seen with the lactatebread.

The improved glycemia, following ingestion ofbread containing lactic acid, is in accordance with thelow metabolic response recently reported for starch ina sourdough fermented pumpernickel bread containing mainly acetic acid (pH 4.5, SÂ°31.1, 29.7

mmol/100 g acetic acid on dry weight basis) (Liljebergand BjÃ¶rck 1994). In addition to the effect observedwith acetic acid (Liljeberg and BjÃ¶rck 1994) or Na-propionate (Tedesco et al. 1991), a reduced glycemiahas also been found in healthy subjects after consumption of fermented vegetables (Torsdottir et al.1992) and lettuce dressed with vinegar (Brighenti andTestolin 1991).

tiM

12O

"Time <min)

T SO

FIGURE 3 Mean satiety scores obtained in healthy subjects following ingestion of breakfast meals with differentbread products. Values are means, n = 11. Error terms aregiven in Table 7.


jn.nutrition.orgD

ownloaded from



TABLE 6

Hydrolysis rate indices of the different breads^

HydrolysisrateindexWhole-meal

breadPlus sourdoughPlus lacticacidPlus

Ca-lactatePlus Na-propionatePlus Na-propionate, high concentration

White wheat bread95.5

Â±2.2C85.5 Â±1.4d80.6 Â±1.3d113.4

Â±3.7a101.0 Â±1.3b103.5 Â±2.7b100bc

Values are means Â±SEM,n - 11. Values not sharing the sameletters are significantly different (P < 0.05).

According to Todesco et al. (1991), the suggestedmechanism for an improved glucose tolerance by Na-propionate is an inhibition of amylolytic enzymes.Consequently, a reduced rate of starch hydrolysis bysalivary amylase was noted in vitro. In contrast, asjudged from in vitro results in the enzymic modelused in the present study, the rate of starch hydrolysiswas not affected by Na-propionate. The hydrolysisrate index values of the WMB plus Na-propionate andWMB plus high Na-propionate were found to be 101and 104, respectively. In fact, when using the same invitro procedure as described by Todesco et al. (1991),we failed to show any inhibition of amylase activityby Na-propionate. However, the bread products containing lactic acid (sourdough and lactic acid), significantly reduced the rate of starch hydrolysis(hydrolysis rate index 86 and 81, respectively), andthe lowered hydrolysis rate index values obtainedwith these bread products can probably be explainedby an inhibition of amylolytic enzymes.

In previous experiments, the in vitro rate of starchhydrolysis obtained with the system used in thepresent study was found to correlate well with blood

glucose responses in healthy subjects (Granfeldt et al.1992). Consequently, a correlation coefficient of 0.826(P < 0.00001) was obtained when plotting glycÃ©mieindex vs. hydrolysis rate index for an importantnumber of cereal and legume products (glycÃ©mieindex = 0.862-hydrolysis rate index + 8.198) (Granfeldt1994). The lack of relationship between glycÃ©mieindex and hydrolysis rate index in the case of thebreads with added Ca-lactate or Na-propionate is thusin contrast to previous experiences with starchyproducts, and suggests that the cause for a reductionin glycÃ©mie and insulinemic responses with thesebread products is related to a mechanism other than areduced rate of starch digestion. When glycÃ©mieindexwas calculated from hydrolysis rate index in the sourdough bread and WMB plus lactic acid, predictedglycÃ©mie indices of 82 and 78, respectively, wereobtained. This 20% reduction is in accordance withthe lowering of the 45-min postprandial glucose areaswith these products. These results suggest that thepresence of lactic acid did reduce the rate of starchdigestion also in the gastrointestinal tract.

The fact that no inhibitory effect, but rather anincrease in hydrolysis rate index was noted with theCa-lactate-containing product, might be related to animproved a-amylase activity in the presence of Ca2+

(Greenwood and Milne 1968). It is not knownwhether this also occurs in vivo. To our knowledge,no data are available on the potential effects of Ca-lactate on the rate of gastric emptying. However, asjudged from the lack of effect of Ca-lactate on postprandial glucose or insulin responses, the overall rateof starch uptake was not affected.

The observed in vitro inhibition of enzyme activityin lactic acid-baked breads (pH 4.0-4.2) is in conflictwith results reported for a sourdough fermented pumpernickel bread containing mainly acetic acid (pH4.5). Hence, the starch in that pumpernickel breadwas as rapidly digested as starch in a white wheatbread (Liljeberg and BjÃ¶rck 1994). However, despitethe lack of effect on in vitro starch hydrolysis rate,the pumpernickel bread containing acetic acid flattened the postprandial glucose and insulin response

TABLE 7

Satiety scores obtained in healthy humans after different bread meals1

Satiety score

Whole-mealbreadPlussourdoughPlus


Na-propionate, high concentration0

min-5.4

Â±0.3-4.9Â±0.4-5.2Â±0.7-4.6Â±0.3-4.8Â±0.5-4.8Â±0.415

min4.2

Â±0.74.2Â±0.74.1Â±0.94.8Â±0.54.5Â±0.94.9Â±1.045

min3.1

Â±0.6a3.5Â±0.5ab4.5Â±0.6ab4.0Â±0.5ab3.8Â±0.6b4.7Â±0.9b95

min1.2

Â±0.7a1.9Â±0.4a1.9Â±0.9ab2.7Â±0.7ab2.5Â±0.6ab3.2Â±0.8b120

min-1.1

Â±0.9a0.4Â±0.4ab-0.3Â±0.8ab1.1Â±0.9b1.4Â±0.7b0.7Â±1.0ab180

min-3.4

Â±l.lab-3.1Â±0.7a-3.1Â±0.9ab-3.0Â±0.9ab-1.8Â±0.6b-2.2Â±l.lab



jn.nutrition.orgD

ownloaded from



TABLE 8

Satiety areas and acceptability scores for the bread products1

Whole-meal breadPlus sourdoughPlus lactic acidPlus Ca-lactatePlus Na-propionatePlus Na-propionate, high

concentrationSatiety

area316.2

Â±76.7a317.2 Â±42.5ab393.9 Â±62.5ab425.9 Â±75.4ab409.7 Â±72.9"b510.7

Â±89.8bAcceptability

score0.8

Â±0.4a-1.9 Â±0.7C-1.4 Â±O.T**

0.1 Â±0.4ab-0.4 Â±0.7abc-1.2

Â±0.6^

Values are means Â±SEM,n = 11. Values not sharing the sameletters are significantly different (P < 0.05].

curves in healthy subjects. A further in vitro evaluation of white wheat bread products has shown thatadded lactic and citric acid reduced the rate ofamylolysis, whereas no inhibition of amylase activitywas found with added acetic or propionic acid (unpublished data). These results indicate that some acidsmay act as amylase inhibitors whereas others do not.

In parallel with the evaluation of metabolicresponses, satiety was studied during 3 h followingthe test meals. All bread products containing organicacid or salt received higher satiety areas under thecurves (0-180 min) than the reference product.However, this tendency only reached statistical significance in the case of a high concentration of Na-

propionate. This may indicate an effect on gastricemptying. In several studies (Hunt and Knox 1969 and1972, Hunt and Pathak 1960), it has been shown thatorganic acids slow gastric emptying in humans,whereas the slowing effects of most salts were foundto be of minor importance (Hunt and Knox 1969,Hunt and Pathak 1960). However, the calcium andsodium salts of lactic and propionic acids, respectively, were not included. In the present study, theobservations with bread products baked with Na-propionate or Ca-lactate were not consistent. Consequently, an improved glycemia was registered withthe breads with added Na-propionate, whereas noeffect was seen with WMB plus Ca-lactate. Propionic

acid is one of the short chain fatty acids producedduring fermentation of undigested carbohydrates inthe colon. It has been suggested that propionic acidhas some regulating functions on hepatic carbohydrate and lipid metabolism (Thorburn et al. 1993). ANa-propionate supplemented diet has also beenshown to lower the hepatic total cholesterol level inrats (unpublished data). Consequently, it is likely thatNa-propionate is metabolized in the liver. It is possible that consumption of Na-propionate-baked

breads also may improve glycemia by influencinghepatic regulation.

The propionate bread was accepted to a higherextent than the lactic acid products. To improvepalatability while still retaining effects on metabolicresponses, it may be preferable to bake bread productswith Na-propionate, or Na-propionate in combinationwith organic acid.

All bread products had a similar content of in vitroresistant starch (1.3-2.1 g/100 g starch basis). Moststudies have shown equally low levels in flour-basedbread (Englyst et al. 1992, Liljeberg and BjÃ¶rck 1994,Liljeberg et al. 1992). Recently, a high content ofresistant starch was found in a commercial sourdoughfermented pumpernickel bread (8.1 g/100 g starchbasis) (Liljeberg and BjÃ¶rck 1994). The results fromthe present study may exclude the effect of organicacids per se on rÃ©trogradation of amylose. Probably,the high content of in vitro resistant starch was dueto the long baking time at low temperature, commonly used for this type of bread.

Results from the present study show that breadproducts baked with lactic acid, formed during sourdough fermentation or added, improved glucosetolerance. Sourdough baking is an old, traditionallyused process and its effects on dough rising, breadtaste, flavor, texture and shelf-life are well established(LÃ¶nner 1988). The information in the present paperconcerning improved nutritional characteristics ofstarch in sourdough bread is, however, new.

Favorably low glucose and insulin responses werealso found with the bread products containing Na-propionate. Obviously, the metabolic effects arerelated to the concentration of added Na-propionate,

i.e., the improved glycemia was more pronouncedwith high Na-propionate. The suggested mechanismfor slowly absorbed starch in the presence of lacticacid is an inhibition of the amylolytic enzymes, asjudged from in vitro results. However, observationswith the product baked with a high concentration ofNa-propionate, e.g. delayed glucose and insulin peakvalues, prolonged satiety and the lack of effect on invitro starch hydrolysis rate, together suggest a mechanism related to a delay in gastric emptying. Weconclude that the beneficial effects of organic acids onpostprandial glycemia could motivate a renewed interest in sourdough baking and other fermentationprocesses. Addition of lactic acid or Na-propionatecould be explored in the baking process for development of lente bread products. The metaboliceffects of corresponding salts could be helpful in improving starch properties in products with less acidiccharacteristics.

ACKNOWLEDGMENTS

We thank Anders Drews for his invaluable assistance. The bread products were skillfully baked byArne Bjager, Nord Mills AB (MalmÃ¶, Sweden).


jn.nutrition.orgD

ownloaded from



LITERATURE CITED

Asp, N.-G., Johansson, C.-G., Hallmer, H. & SiljestrÃ¶m, M. (1983)Rapid enzymic method for assay of insoluble and soluble fiber. J.Agrie. Food Chem. 31: 476-482.

Brand, I., Colagiuri, S., Grossman, S., Allen, A., Roberts, D. &Truswell, S. (1991) Low glycÃ©mieindex foods improve longtermglycÃ©miecontrol in NIDDM. Diabetes Care 14: 95-101.

Brighenti, F. & Testolin, G. (1991) Effects of a limited dose ofvinegar on blood glucose and acetate responses to a mixed mealin healthy subjects. Poster presented at the ninth internationalsymposium on diabetes and nutrition, Diabetes and NutritionStudy Group of the EASD, June 13-15, Uppsala, Sweden.

Chapman, L. D. & Wigfield, R. (1970) Rating scales in consumerresearch. Food Manuf. 45: 59-62.

Collier, G. R., Wolever, T.M.S., Wong, G. S. &. Josse, R. G. (1986)Prediction of glycÃ©mieresponse to mixed meals in noninsulin-dependent diabetic subjects. Am. I. Clin. Nutr. 44: 349-352.

Conover, W.}. (1980) Practical nonparametric statistics. John Wiley& Sons, New York, NY.

Ducimetiere, P., Eschwege, E., Papox, L., Richard, J. L., Claude, J. R.& Rosselin, G. (1980) Relationship of plasma insulin levels tothe incidence of myocardial infarction and coronary heartdisease mortality in a middle-aged population. Diabetologia 19:205-210.

Englyst, H. N., Kingman, S. M. & Cummings, f. H. (1992) Classification and measurement of nutritionally important starchfraction. Eur. J. Clin. Nutr. 46 (Suppl.): S33-S50.

Granfeldt, Y. (1994) Food Factors Affecting Metabolic Responses toCereal Products. Doctoral thesis, University of Lund, Lund,Sweden.

Granfeldt, Y. & BjÃ¶rck,I. (1991) GlycÃ©mieresponse to starch inpasta: a study of mechanism of limited enzyme availability. }.Cereal Sci. 14: 47-61.

Granfeldt, Y., BjÃ¶rck,I., Drews, A. & Tovar, J. (1992) An in vitroprocedure based on chewing to predict metabolic response tostarch in cereal and legume products. Eur. J. Clin. Nutr. 46:649-660.

Granfeldt, Y., Liljeberg, H., Drews, A., Newman, R. & BjÃ¶rck,I.(1994) Glucose and insulin responses to barley products: influence of food structure and amylose-amylopectin ratio. Am. J.Clin. Nutr. 59: 1075-1082.

Greenwood, C. T. & Milne, E. A. (1968) Studies on starch-degradingenzymes. A comparison of a-amylases from different sources:their properties and action patterns. Die StÃ¤rke20: 139-150.

Haber, G. B., Heaton, K. W., Murphy, D. & Burroughs, L. F. (1977)Depletion and disruption of dietary fibre. Effects on satiety,plasma-glucose, and serum-insulin. Lancet 1: 679-682.

Holm, f., BjÃ¶rck,I., Drews, A. & Asp, N.-G. (1986) A rapid methodfor the analysis of starch. Starch/Staerke 38: 224-226.

Holt, S., Brand, }., Soveny, C. & Hansky, J. (1992) Relationship of

satiety to postprandial glycaemic, insulin and cholecystokininresponses. Appetite 18: 129-141.

Hostettler, F., Borei, E. & Devel, H. (1951) Ãœberdie reduktion der3,5-di-nitrosalicylsÃ¤ure durch zucker. Helvet. Chim. Acta. 34:2132-2139.

Hunt, J. N. &. Knox, M. T. (1969) The slowing of gastric emptyingby nine acids. J. Physiol. 201: 161-179.

Hunt, J. N. &. Knox, M. T. (1972) The slowing of gastric emptyingby four strong acids and three weak acids. J. Physiol. 222:187-208.

Hunt, I. N. & Pathak, ]. D. (1960) The osmotic effect of somesimple molecules and ions in gastric emptying. J. Physiol. 154:254-269.

Jenkins, D.J.A., Wolever, T.M.S., Collier, G. R., Ocana, A., Rao, A.V., Buckley, G., Lam, Y., Mager, A. & Thompson, L. U. (1987)Metabolic effects of a low-glycemic-index diet. Am. J. Clin.Nutr. 46: 968-975.

Jenkins, D.J.A., Wolever, T.M.S. & Jenkins, A. L. (1988) Starchyfoods and glycÃ©mieindex. Diabetes Care 11: 149-159.

Liljeberg, H. & BjÃ¶rck,I. (1994) Bioavailability of starch in breadproducts. Postprandial glucose and insulin responses in healthysubjects and in vitro resistant starch content. Eur. J. Clin. Nutr.48: 151-163.

Liljeberg, H., Granfeldt, Y. & BjÃ¶rck,l. (1992) Metabolic responsesto starch in bread containing intact kernels versus milled flour.Eur. J. Clin. Nutr. 46: 561-575.

LÃ¶nner, C. (1988) Starter Cultures for Rye Sour Doughs. Characteristics and Functions of Lactic Acid Bacteria. Doctoral thesis,University of Lund, Lund, Sweden.

LÃ¶nner,C. & PrevÃ©-Akesson,K. (1988) Acidification properties oflactic acid bacteria in rye sour doughs. Food Microbiol. 5: 43-58.

SiljestrÃ¶m, M., Eliasson, A.-C. & BjÃ¶rck,I. (1989) Characterizationof resistant starch from autoclaved wheat starch. Starch/Staerke41: 147-151.

Spicher, G. & LÃ¶nner,C. (1985) Die mikroflora des Sauerteiges XXI.Die in Sauerteigen Schwedischer bÃ¤ckereien vorkommendenLactobacillen. Z. Lebensm. Unters. Forsch. 181: 9-13.

Thomas, D. E., Brotherhood, J. R. &. Brand, J. C. (1991) Carbohydrate feeding before exercise: effect of glycÃ©mieindex. Int. J.Sports Med. 12: 180-186.

Thorburn, A., Muir, J. & Proietto, J. (1993) Carbohydrate fermentation decreases hepatic glucose output in healthy subjects.Metabolism 42: 780-785.

Todesco, T., Rao, A. V., Bosello, O. & Jenkins, D.J.A. (1991) Propi-onate lowers blood glucose and alters lipid metabolism inhealthy subjects. Am. J. Clin. Nutr. 54: 860-865.

Torsdottir, I., Blomqvist, R., Ekman, R., Hagander, B. & Sandberg,A-S. (1992) Fermented or fresh vegetables decrease the postprandial blood glucose and insulin levels in healthy persons.Scand. J. Nutr. 36: 4-7.

Tovar, J., Granfeldt, Y. & BjÃ¶rck,I. (1992) Effect of processing onblood glucose and insulin responses to starch in legumes. J.Agrie. Food Chem. 40: 1846-1851.


jn.nutrition.orgD

ownloaded from


sourdough fermentation or addition of organic acids or

Documents